DOCSLIB.ORG

AGENT MODELING of HUMAN INTERACTION: Stability, Dynamics and Cooperation

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
AGENT MODELING of HUMAN INTERACTION: Stability, Dynamics and Cooperation AGENT MODELING OF HUMAN INTERACTION: Stability, Dynamics and Cooperation Thesis submitted for the degree of \Doctor of Philosophy" By Omer Lev Submitted to the Senate of the Hebrew University of Jerusalem August 2015 This work was carried out under the supervision of: Jeffrey S. Rosenschein iv Abstract This work models human behavior and interactions, particularly in domains where there are results indicating that game theoretic solutions diverge from real world dynamics. We do so by extending existing models to incorporate better heuristics and approximations of how people behave and interact. We begin with voting and social-choice where, as the Gibbard-Satterthwaite theorem states, no reasonable voting rule is strategyproof; i.e., there will always be scenarios in which voters are better off mis-reporting their pref- erences and voting strategically. This suggests that in order to understand election results, we should look at Nash equilibria { states where no voter wishes to change their vote. However, not only are there many equilibria, but a significant number of them would never be the outcome of real world votes. We propose several enhancements to the standard voting model to handle these problems: assumptions on voters' nature (a slight preference of voters to vote for what they truly believe); assumptions on voters' dynamics (they participate in an iterative process, in which each voter changes their preference if they believe it will change the outcome); and a more complex model where voters have a view of \plausible" election outcomes based on some notion they have of how others will vote (e.g., a poll). In addition to theoretical proofs and characterizations, we also provide simulation results indicating that these models give us more realistic outcomes. We continue with a similar problem in which Nash Equilibrium, as a so- lution concept, does not correspond to real world outcomes. All-pay auctions model various human endeavors such as the development of medical drugs v vi or crowdsourcing, in which many participants exert effort, but only one wins (e.g., a drug patent) while the rest lose their investment. A naive analysis indicates that no one should participate in such auctions, as the expected profit is zero. We show that assuming players collude (i.e., cooperate with- out other players' knowledge) increases the colluders' expected profit, and in some cases increases the expected profit of the non-colluders as well, due to their ignorance of the collusion. We further address the same problem via a different path, and show that assuming auction participants are not fixed, but rather have a certain probability of participating (as in, for example, crowdsourcing), the expected profit is positive as well. We conclude with two different approaches to using graphs and networks for modeling interactions. Using cooperative game theory we analyze weakest link games, in which a path between a source and a target is valued as the weight of the lightest edge in the path. We show algorithms for finding stable solutions, and the NP-completeness of finding optimal coalition structure (as well as an appropriate approximation algorithm). In particular, we focus on cost of stability calculations, which determine how much more needs to be given to the game scenario to ensure its stability. We then turn to utilizing the graph in a different manner, as representing a social graph, to parts of which we wish to give group recommendations (e.g., what restaurant shall a group of friends go to, or what game should they play together). We pursue an axiomatic approach, in which we specify desirable features, showing that some are incompatible, while some result in a unique recommendation algorithm. Improved models enable a better understanding of humans, capturing in each setting the most salient features that allow us to better represent interactions. Second, they enable us to design agents that react and interact with people, understanding, albeit to a limited degree, how people behave in certain settings. Finally, this enables better mechanism design, giving system designers better models and tools to construct their solutions. Contents 1 Introduction 1 1.1 The Context of Modeling Human Behavior . 1 1.2 Decision Making and Elections . 4 1.3 All-Pay Auctions . 6 1.4 Networks { Human and Otherwise . 6 1.5 A Note on Papers . 9 I Decision-Making and Elections 11 2 Decision-Making Overview & Preliminaries 13 2.1 Preliminaries and Definitions . 15 2.1.1 Voting Rules . 17 2.1.2 Voters: Truth-Bias and Lazy-Bias . 19 2.1.3 Iterative Voting . 20 3 The Truth(-Bias) is Out There: Truth-Biased Voters 23 3.1 Introduction . 23 3.1.1 Related Work . 24 3.2 Simulation Methodology . 26 3.3 Simulation Results . 28 3.4 Analytical Results . 32 3.4.1 Veto Voting Rule . 33 CONTENTS 3.4.2 A Constructive Algorithm . 34 3.4.3 k-Approval . 35 3.5 Summary . 36 4 Don't stop 'Til You Get Enough: Iterative Voting 39 4.1 Introduction . 39 4.1.1 Related Work . 41 4.2 Tie-Breaking Rules . 43 4.3 Voting Rules . 43 4.3.1 Veto . 43 4.3.2 Other Scoring Rules . 44 4.4 Voter Types . 44 4.4.1 Regular Voters . 44 4.4.2 Truth-Biased Voters . 44 4.4.3 Lazy-Biased Voters . 47 4.5 Summary . 48 5 Think Local, Act Global: Local Dominance Voting 53 5.1 Introduction . 53 5.1.1 Related Work . 55 5.2 The Local Dominance Model . 56 5.2.1 Locality . 56 5.2.2 Dominance . 58 5.2.3 Truth and Lazy Bias . 58 5.2.4 Generality . 60 5.3 Iterative Convergence . 60 5.4 Truth-Bias and Lazy-Bias . 62 5.5 Simulation Results . 62 5.6 Summary . 66 CONTENTS II All-Pay Auctions 69 6 All-Pay Auction Overview & Preliminaries 71 6.1 Preliminaries and Definitions . 73 6.1.1 Bids . 74 6.1.2 Bidders . 75 6.1.3 Auctioneers . 76 7 Come Together: Mergers and Collusions in All-Pay Auctions 79 7.1 Introduction . 79 7.1.1 Related Work . 80 7.2 Mergers . 82 7.3 Collusions . 83 7.3.1 A Single Group of Colluders . 84 7.3.2 Optimal Number of Colluders . 85 7.3.3 Auctioneer's Profits . 86 7.3.4 Social Welfare . 86 7.3.5 Other Equilibria . 87 7.4 Response To Collusion . 88 7.4.1 A Player Aware Of Collusion . 88 7.4.2 Several Groups Of Colluders . 89 7.5 Summary . 90 8 \Showing Up is 80% of Life": Participation in All-Pay Auc- tions 93 8.1 Introduction . 93 8.1.1 Related Work . 94 8.2 Equilibrium . 95 8.3 Profits . 97 8.4 Special Case: A Single Participation Probability . 100 8.5 Summary . 101 CONTENTS III Networks 103 9 Networks Overview 105 10 \You are the Weakest Link, Goodbye!": Weakest Link Games107 10.1 Introduction . 107 10.1.1 Related Work . 109 10.2 Preliminary: Cooperative Game Theory . 111 10.2.1 Stability Concepts . 112 10.2.2 Coalition Structures . 114 10.3 Weakest Link Game Definition . 114 10.4 Core and Least-Core . 116 10.5 Cost of Stability and Series and Parallel Compositions . 119 10.6 Optimal Coalition Structures . 121 10.7 Summary . 121 11 Where Do We Want to Go Today?: Group Recommenda- tions 123 11.1 Introduction . 123 11.1.1 Related Work . 125 11.2 The Group Recommendation Model . 126 11.3 The Axioms . 128 11.3.1 The Basic Axioms . 128 11.3.2 Group Power Axioms . 130 11.3.3 Influence Structure Axioms . 132 11.3.4 Independence of Axioms from Each Other . 133 11.4 An Impossibility Result . 134 11.5 A Recommendation System that Works . 135 11.6 Summary . 137 12 When All is Said and Done. 139 Bibliography 143 CONTENTS IV Appendices 175 A Proofs of Part I 177 A.1 Chapter 3 . 177 A.2 Chapter 4 . 186 A.3 Chapter 5 . 203 B Proofs of Part II 209 B.1 Chapter 7 . 209 B.2 Chapter 8 . 214 C Proofs of Part III 217 C.1 Chapter 10 . 217 C.2 Chapter 11 . 224 D Information on Local Dominance Simulations 233 D.1 Distributions . 234 D.2 Collected Variables . 235 CONTENTS Lord, what fools these mortals be! William Shakespeare, A Midsummer Night's Dream Chapter 1 Introduction 1.1 The Context of Modeling Human Behavior Understanding and analyzing human behavior has been attempted since the dawn of human societies, as rulers (and their nascent bureaucracies) have struggled to understand how they can implement their policies with little resistance, and how they can motivate people towards their desired goals. Naturally, philosophers have ventured to answer this, and diverse schools of thought { from Plato and his Greek contemporaries to Confucius in China { have tried to understand human behavior and develop \mechanisms" that would change this behavior more to their liking. These attempts { and the attempts in the millennia that followed { were mostly based on an effort to generalize human behavior from people's ob- served actions in particular times and places. This was usually done through a moral prism, often influenced by the theological views of the time. While the philosophical efforts and debates continued unabated, the ad- vent of the modern age during the late 18th century and early 19th century, made the need for a generalized analysis of human behavior even more acute and necessary [103, 234, 221]: 1 2 CHAPTER 1. INTRODUCTION • The large state: Countries began increasing in size, and along with the speeding of communication, this meant many decision were taken in central bureaucracies, where detailed knowledge of local conditions was limited, to a large extent.
Load more

Recommended publications
  • Efficient Algorithms for Computing Approximate Equilibria in Bimatrix
    Efficient Algorithms for Computing Approximate Equilibria in Bimatrix, Polymatrix and Lipschitz Games Thesis submitted in accordance with the requirements of the University of Liverpool for the degree of Doctor in Philosophy by Argyrios Deligkas July 2016 Abstract In this thesis, we study the problem of computing approximate equilibria in sev- eral classes of games. In particular, we study approximate Nash equilibria and approximate well-supported Nash equilibria in polymatrix and bimatrix games and approximate equilibria in Lipschitz games, penalty games and biased games. We construct algorithms for computing approximate equilibria that beat the cur- rent best algorithms for these problems. In Chapter 3, we present a distributed method to compute approximate Nash equilibria in bimatrix games. In contrast to previous approaches that analyze the two payoff matrices at the same time (for example, by solving a single LP that combines the two players' payoffs), our algorithm first solves two independent LPs, each of which is derived from one of the two payoff matrices, and then computes an approximate Nash equilibrium using only limited communication between the players. In Chapter 4, we present an algorithm that, for every δ in the range 0 < δ ≤ 0:5, finds a (0:5+δ)-Nash equilibrium of a polymatrix game in time polynomial in 1 the input size and δ . Note that our approximation guarantee does not depend on the number of players, a property that was not previously known to be achievable for polymatrix games, and still cannot be achieved for general strategic-form games. In Chapter 5, we present an approximation-preserving reduction from the problem of computing an approximate Bayesian Nash equilibrium (-BNE) for a two-player Bayesian game to the problem of computing an -NE of a polymatrix game and thus show that the algorithm of Chapter 4 can be applied to two-player Bayesian games.
    [Show full text]
  • Direct Tests of Cumulative Prospect Theory⇤
    Direct Tests of Cumulative Prospect Theory⇤ B. Douglas Bernheim † Charles Sprenger‡ Stanford University and NBER UC San Diego First Draft: December 1, 2014 This Version: November 8, 2016 Abstract Cumulative Prospect Theory (CPT), the leading behavioral account of decision making under uncertainty, assumes that the probability weight applied to a given outcome depends on its ranking. This assumption is needed to avoid the violations of dominance implied by Prospect Theory (PT). We devise a simple test involving three-outcome lotteries, based on the implication that compensating adjustments to the payo↵s in two states should depend on their rankings compared with payo↵s in a third state. In an experiment, we are unable to find any support for the assumption of rank dependence. We separately elicit probability weighting functions for the same subjects through conventional techniques involving binary lotteries. These functions imply that changes in payo↵rank should change the compensating adjustments in our three-outcome lotteries by 20-40%, yet we can rule out any change larger than 7% at standard confidence levels. Additional tests nevertheless indicate that the dominance patterns predicted by PT do not arise. We reconcile these findings by positing a form of complexity aversion that generalizes the well-known certainty e↵ect. JEL classification: D81, D90 Keywords:ProspectTheory,CumulativeProspectTheory,RankDependence,CertaintyEquiv- alents. ⇤We are grateful to Ted O’Donoghue, Colin Camerer, Nick Barberis, Kota Saito, and seminar participants at Cornell, Caltech, and Gary’s Conference (UC Santa Barbara) for helpful and thoughtful discussions. Fulya Ersoy, Vincent Leah-Martin, Seung-Keun Martinez, and Alex Kellogg all provided valuable research assistance.
    [Show full text]
  • Judgment Under Uncertainty: Heuristics and Biases Author(S): Amos Tversky and Daniel Kahneman Source: Science, New Series, Vol
    Judgment under Uncertainty: Heuristics and Biases Author(s): Amos Tversky and Daniel Kahneman Source: Science, New Series, Vol. 185, No. 4157, (Sep. 27, 1974), pp. 1124-1131 Published by: American Association for the Advancement of Science Stable URL: http://www.jstor.org/stable/1738360 Accessed: 15/04/2008 14:50 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=aaas. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We enable the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected]. http://www.jstor.org occupation from a list of possibilities (for example, farmer, salesman, airline pilot, librarian, or physician)? How do people order these occupations from most to least likely? In the representa- Judgment under Uncertainty: tiveness heuristic, the probability that Steve is a librarian, for example, is Heuristics and Biases assessed by the degree to which he is representative of, or similar to, the stereotype of a librarian.
    [Show full text]
  • THE HUMAN SIDE of MECHANISM DESIGN a Tribute to Leo Hurwicz and Jean-Jacque Laffont
    THE HUMAN SIDE OF MECHANISM DESIGN A Tribute to Leo Hurwicz and Jean-Jacque Laffont Daniel McFadden Department of Economics University of California, Berkeley December 4, 2008 [email protected] http://www.econ.berkeley.edu/~mcfadden/index.shtml ABSTRACT This paper considers the human side of mechanism design, the behavior of economic agents in gathering and processing information and responding to incentives. I first give an overview of the subject of mechanism design, and then examine a pervasive premise in this field that economic agents are rational in their information processing and decisions. Examples from applied mechanism design identify the roles of perceptions and inference in agent behavior, and the influence of systematic irrationalities and sociality on agent responses. These examples suggest that tolerance of behavioral faults be added to the criteria for good mechanism design. In principle-agent problems for example, designers should consider using experimental treatments in contracts, and statistical post-processing of agent responses, to identify and mitigate the effects of agent non-compliance with contract incentives. KEYWORDS: mechanism_design, principal-agent_problem, juries, welfare_theory JEL CLASSIFICATION: D000, D600, D610, D710, D800, C420, C700 THE HUMAN SIDE OF MECHANISM DESIGN A Tribute to Leo Hurwicz and Jean-Jacque Laffont Daniel McFadden1 1. Introduction The study of mechanism design, the systematic analysis of resource allocation institutions and processes, has been the most fundamental development
    [Show full text]
  • Behavioral Economics in Context Applications for Development, Inequality & Discrimination, Finance, and Environment
    Behavioral Economics In Context Applications for Development, Inequality & Discrimination, Finance, and Environment By Anastasia C. Wilson An ECI Teaching Module on Social and Environmental Issues in Economics Global Development Policy Center Boston University 53 Bay State Road Boston, MA 02155 bu.edu/gdp Economics in Context Initiative, Global Development Policy Center, Boston University, 2020. Permission is hereby granted for instructors to copy this module for instructional purposes. Suggested citation: Wilson, Anastasia C. (2020) “Behavioral Economics In Context: Applications for Development, Inequality & Discrimination, Finance, and Environment.” An ECI Teaching Module on Social and Economic Issues, Economics in Context Initiative, Global Development Policy Center, Boston University, 2020. Students may also download the module directly from: http://www.bu.edu/eci/education-materials/teaching-modules/ Comments and feedback from course use are welcomed: Economics in Context Initiative Global Development Policy Center Boston University 53 Bay State Road Boston, MA 02215 http://www.bu.edu/eci/ Email: [email protected] NOTE – terms denoted in bold face are defined in the KEY TERMS AND CONCEPTS section at the end of the module. BEHAVIORAL ECONOMICS IN CONTEXT TABLE OF CONTENTS 1. INTRODUCTION ........................................................................................................ 4 1.1 The History and Development of Behavioral Economics .......................................................... 5 1.2 Behavioral Economics Toolkit:
    [Show full text]
  • Bounded Rationality: Models of Fast and Frugal Inference
    Published in: Swiss Journal of Economics and Statistics, 133 (2/2), 1997, 201–218. © 1997 Peter Lang, 0303-9692. Bounded Rationality: Models of Fast and Frugal Inference Gerd Gigerenzer1 Max Planck Institute for Psychological Research, Munich, Germany Humans and other animals need to make inferences about their environment under constraints of limited time, knowledge, and computational capacities. However, most theories of inductive infer- ences model the human mind as a supercomputer like a Laplacean demon, equipped with unlim- ited time, knowledge, and computational capacities. In this article I review models of fast and frugal inference, that is, satisfi cing strategies whose task is to infer unknown states of the world (without relying on computationaly expensive procedures such as multiple regression). Fast and frugal inference is a form of bounded rationality (Simon, 1982). I begin by explaining what bounded rationality in human inference is not. 1. Bounded Rationality is Not Irrationality In his chapter in John Kagel and Alvin Roth’s Handbook of Experimental Economics (1995), Colin Camerer explains that “most research on individual decision making has taken normative theo- ries of judgment and choice (typically probability rules and utility theories) as null hypotheses about behavior,” and has labeled systematic deviations from these norms “cognitive illusions” (p. 588). Camerer continues, “The most fruitful, popular alternative theories spring from the idea that limits on computational ability force people to use simplifi ed procedures or ‘heuristics’ that cause systematic mistakes (biases) in problem solving, judgment, and choice. The roots of this approach are in Simon’s (1955) distinction between substantive rationality (the result of norma- tive maximizing models) and procedural rationality.” (p.
    [Show full text]
  • Rationally Biased Learning
    Rationally Biased Learning Michel De Lara CERMICS, Ecole des Ponts, Marne-la-Vall´ee, France Abstract Humans display a tendency to pay more attention to bad outcomes, often in a disproportionate way relative to their statistical occurrence. They also display euphorism, as well as a preference for the current state of affairs (status quo bias). Based on the analysis of optimal solutions of infinite horizon stationary optimization problems under imperfect state observation, we show that such human perception and decision biases can be grounded in a form of rationality. We also provide conditions (boundaries) for their possible occurence and an analysis of their robustness. Thus, biases can be the product of rational behavior. Keywords: pessimism bias, status quo bias, euphorism bias, probability overestimation, optimal behavior, imperfect state information. 1. Introduction When we perceive sounds, we overestimate the change in level of rising level tones relative to equivalent falling level tones (Neuhoff (1998)). When we assess pros and cons in decision making, we weigh losses more than gains (Kahneman and Tversky (1979)). We are more frightened by a snake or a arXiv:1709.02256v2 [cs.AI] 23 Oct 2020 spider than by a passing car or an electrical shuffle. Such human assessments are qualified of biases, because they depart from physical measurements or objective statistical estimates. Thus, there is “bias” when a behavior is not aligned with a given “rationality benchmark” (like expected utility theory), as documented in the “heuristics and biases” literature (Kahneman et al. (1982); Gilovich et al. (2002)). However, if such biases are found consistently in human behavior, they must certainly have a reason.
    [Show full text]
  • Behavioral Mechanism Design: Optimal Crowdsourcing Contracts and Prospect Theory
    Behavioral Mechanism Design: Optimal Crowdsourcing Contracts and Prospect Theory David Easley, Cornell University Arpita Ghosh, Cornell University Incentive design is more likely to elicit desired outcomes when it is derived based on accurate models of agent behavior. A substantial literature in behavioral economics, however, demonstrates that individuals systematically and consistently deviate from the standard economic model—expected utility theory—for decision-making under uncertainty, which is at the core of the equilibrium analysis necessary to facilitate mechanism design. Can these behavioral biases—as modeled by prospect theory [Kahneman and Tversky 1979]—in agents’ decision-making make a difference to the optimal design of incentives in these environments? In this paper, we explore this question in the context of markets for online labor and crowdsourcing where workers make strategic choices about whether to undertake a task, but do not strategize over quality conditional on participation. We ask what kind of incentive scheme—amongst a broad class of contracts, including those observed on major crowdsourcing platforms such as fixed prices or base payments with bonuses (as on MTurk or oDesk), or open-entry contests (as on platforms like Kaggle or Topcoder)—a principal might want to employ, and how the answer to this question depends on whether workers behave according to expected utility or prospect theory preferences. We first show that with expected utility agents, the optimal contract—for any increasing objective of the principal— always takes the form of an output-independent fixed payment to some optimally chosen number of agents. In contrast, when agents behave according to prospect theory preferences, we show that a winner-take-all contest can dominate the fixed-payment contract, for large enough total payments, under a certain condition on the preference functions; we show that this condition is satisfied for the parameters given by the literature on econometric estimation of the prospect theory model [Tversky and Kahneman 1992; Bruhin et al.
    [Show full text]
  • Economics 200C
    Economics 206, Decisions Winter 2007 Vincent Crawford Economics 319, 534-3452, [email protected] Organization: Economics 206 meets from 11:00-12:20 on Mondays and Wednesdays, January 8 through March 14 (university holidays January 15 and February 19 excepted), in Economics 210. My office hours throughout the quarter will be Wednesdays from 2:00-3:00 or by appointment. The course website is http://weber.ucsd.edu/~vcrawfor/econ206.html. The first half (approximately) will cover standard topics in the traditional theory of decisions. The second half will cover topics in behavioral decision theory. Instead of a final exam there will be two take-home mini-exams, one for each half. You must work on these individually, without consulting anyone but me. The first mini-exam will be posted on the website by 4 p.m. Wednesday, February 7, and due by email or in the course mailbox in Economics Student Services by 4 p.m. Friday, February 9. The second mini- exam will be posted by 4 p.m. Wednesday, March 14, and due by email or in the course mailbox by 4 p.m. Friday, March 16 (the last day of classes). The 48-hour time limits are a humanitarian gesture, and should not be binding. The dates are firm except for extreme unforeseeable events, in which case exceptions must be requested as soon as possible. I will also post optional problem sets, which should be good practice for the exams, before each segment. The second exam will include a flexible essay question on behavioral decision theory, which will be posted by the second half.
    [Show full text]
  • 1 Economic Theory and Contemporary Social Problems
    The Human Side of Economic Analysis: Economic Environments and the Evolution of Norms and Preferences Herb ert Gintis Paul Romer Department of Economics Department of Economics University of Massachusetts University of California, Berkeley Amherst, MA 01003 Berkeley, CA 94720 July 15, 1998 1 Economic Theory and Contemporary Social Problems From its inception conventional economic theory was designed to deal with the pro- duction and consumption of marketable goods and services. With the growth of the welfare state and the need for demand management in the p ost World War I I p erio d, economic theory broadened its p ersp ective to include public go o ds and a state sector, considerably improving its ability to deal with the role of government in a ecting eco- nomic stability, growth, and equity. But economic theory to day faces a novel set of so cial problems for which it remains ill-prepared. This research pro ject addresses a key element in the underlying structure of economic theory contributing to its current weakness: its representation of the individual human actor. Conventional economics uses a rational actor mo del that treats the individual as a self-interested utility maximizer with preferences over a narrow range of p ersonal con- sumption items. Preferences in the rational actor mo del are moreover considered as determined outside the economic system and una ected by the individual's economic b ehavior. As we do cument b elow, a considerable body of empirical evidence contra- dicts this view: p eople are motivated by duty
    [Show full text]
  • Adam Smith, Behavioral Economist
    Journal of Economic Perspectives—Volume 19, Number 3—Summer 2005—Pages 131–145 Adam Smith, Behavioral Economist Nava Ashraf, Colin F. Camerer and George Loewenstein n The Wealth of Nations, published in 1776, Adam Smith famously argued that economic behavior was motivated by self-interest. But 17 years earlier in 1759, I Smith had proposed a theory of human behavior that looks anything but self- interested. In his first book, The Theory of Moral Sentiments, Smith argued that behavior was determined by the struggle between what Smith termed the “passions” and the “impartial spectator.” The passions included drives such as hunger and sex, emotions such as fear and anger, and motivational feeling states such as pain. Smith viewed behavior as under the direct control of the passions, but believed that people could override passion-driven behavior by viewing their own behavior from the perspective of an outsider—the impartial spectator—a “moral hector who, looking over the shoulder of the economic man, scrutinizes every move he makes” (Grampp, 1948, p. 317).1 1 A long-standing dispute has raged over whether Adam Smith’s view of human motivation as expressed in The Theory of Moral Sentiments complements or contradicts the view of human motivation expressed in The Wealth of Nations. Although much has been written about “das Adam Smith problem” of reconciling these texts, most modern Smith scholarship asserts that there is no essential contradiction between the texts. As the editors of the Glasgow Edition of the Works and Correspondence of Adam Smith edition of The Theory of Moral Sentiments write (Raphael and Macfie, 1976, p.
    [Show full text]
  • The Origin of Bounded Rationality and Intelligence1
    The Origin of Bounded Rationality and Intelligence1 ANDREW W. LO Charles E. and Susan T. Harris Professor, MIT Sloan School of Management Principal Investigator, MIT Computer Science and Artificial Intelligence Laboratory Director, MIT Laboratory for Financial Engineering Chief Investment Strategist, AlphaSimplex Group, LLC Abstract. Rational economic behavior in which individuals maximize their own self-interest is only one of many possible types of behavior that arise from natural selection. Given an initial population of individuals, each assigned a purely arbitrary behavior with respect to a binary choice problem, and assuming that offspring behave identically to their parents, only those behaviors linked to reproductive success will survive, and less successful behaviors will disappear exponentially fast. This frame- work yields a single evolutionary explanation for the origin of several behaviors that have been observed in organisms ranging from bacteria to humans, including risk-sensitive foraging, risk aversion, loss aversion, probability matching, randomization, and diversification. The key to un- derstanding which types of behavior are more likely to survive is how be- havior affects reproductive success in a given population’s environment. From this perspective, intelligence is naturally defined as behavior that increases the likelihood of reproductive success, and bounds on rational- ity are determined by physiological and environmental constraints. would like to thank Dr. Harriet Zuckerman for inviting me to speak before the American Philosophical Society. I have to say that it is a par- ticular pleasure and privilege for me to be delivering this talk on bounded I rationality and intelligence to the society founded by one of the earliest 1 Read 13 November 2010 as part of the symposium “Current Research in the Social and Behavioral Sciences: Across Disciplinary Lines” and based on joint research with Thomas J.
    [Show full text]